canada: selected issues - international monetary fund · 63 washington mutual inc. usa 8.5 100...
TRANSCRIPT
© 2009 International Monetary Fund May 2009
IMF Country Report No. 09/163
Canada: Selected Issues
This Selected Issues paper for Canada was prepared by a staff team of the International Monetary Fund as background documentation for the periodic consultation with the member country. It is based on the information available at the time it was completed on April 22, 2009. The views expressed in this document are those of the staff team and do not necessarily reflect the views of the government of Canada or the Executive Board of the IMF. The policy of publication of staff reports and other documents by the IMF allows for the deletion of market-sensitive information.
Copies of this report are available to the public from
International Monetary Fund ● Publication Services 700 19th Street, N.W. ● Washington, D.C. 20431
Telephone: (202) 623-7430 ● Telefax: (202) 623-7201 E-mail: [email protected] ● Internet: http://www.imf.org
International Monetary Fund Washington, D.C.
INTERNATIONAL MONETARY FUND
CANADA
Selected Issues Prepared by M. Estevão, R. Duttagupta, E. Tsounta, L. Ratnovski, N. Barrera (all WHD),
J. Kiff (MCM)
Approved by Western Hemisphere Department
April 22, 2009
Contents Page
I. Canadian Banks and the Credit Turmoil............................................................................2 Table 1. Bank Fundamentals and Performance during Turmoil........................................4 Table 2. Bank Fundamentals and Performance: Multivariate Regression Results............5
II. Canadian Residential Mortgage Markets: Boring But Effective? .....................................6 Figure 1. Canadian Bank Mortgage Rates .........................................................................8 Figure 2. Residential Mortgage versus Interest Rate Swap Rates .....................................8
III. The Impact of Global Shocks on Canada—What Do Macro-Financial Linkages Tell Us? ......................................................................................................9 Figure 1. Banks' Lending Standards for Corporations and Real GDP Growth .................9 Figure 2. Impulse Response Functions ............................................................................10 Figure 3. Canadian Real GDP Forecast for 2009 Under Alternative paths for U.S. Real GDP Growth..........................................................................................10 Figure 4. Canadian Real GDP Forecast Under Alternative Conditions for Canadian SLOS......................................................................................................10 Figure 5. Variance Decomposition of Canadian Real GDP Growth ...............................11 Figure 6. Contribution to the Macro-Financial Conditions Index ...................................11
IV. Is the Canadian Housing Market Overvalued? A Tale of Two Regions .........................12 Figure 1. Real House Prices.............................................................................................12 Table 1. Determinants of House Prices in Selected Canadian Provinces ........................13 Figure 2. Canada: House Price Over/Undervaluation......................................................15
V. How Do Commodity Prices Affect Economic Slumps in Commodity Exporters? .........16 Figure 1. Commodity Price Index....................................................................................16 Table 1. Sample of Countries...........................................................................................16 Figure 2. Average Length of Economic Downturns........................................................17 Table 2. Fixed Effect Regressions of Length and Depth of Economic Slump................18
2
I. CANADIAN BANKS AND THE CREDIT TURMOIL1
1. Canadian banks have been relatively resilient to the ongoing credit turmoil. The effect of the turmoil on the Canadian financial system is, so far, milder than in other G7 economies. Interbank money markets remained functional. No injections of public capital into banks were necessary. The resilience appears particularly striking given the close economic and financial links between Canada and the United States.
2. To shed light on this resilience, this chapter searches for key variables explaining Canadian bank performance during the crisis. It considers a sample of large OECD banks and studies how pre-crisis balance-sheet structure affected bank performance during the crisis. The sample includes all large OECD commercial banks (72 institutions with assets in excess of 100 billion euros at the end of 2006). We consider three main fundamentals: the equity-to-assets ratio, the balance-sheet liquidity-to-debt liabilities ratio, and the depository-funding-to-assets ratio. Measures of performance are: equity price decline from January 2007 to January 2009, two binary variables for particularly large equity declines (greater than 70 or 85 percent), and two binary variables for government intervention (undertaken to alleviate significant financial distress).
3. Capital ratios before the crisis were a key determinant of bank performance during the turmoil; and Canadian banks had ample capital. Specifically, most banks with critically low capital at end-2006 later experienced dramatic equity value declines, and many had to be rescued (Table 1). Prior to the crisis, all Canadian banks had capital ratios (equity over assets, a leverage measure) above 4 percent, which has assured their resilience to asset shocks. Interestingly, a large number of currently distressed U.S. banks had relatively high pre-crisis capital, which was nevertheless quickly exhausted through troubled asset exposures and (in some instances) problematic acquisitions.
4. Compared to OECD peers, Canadian banks had slightly above-average balance sheet liquidity. Buffers of highly liquid assets allow banks to bridge temporary cash flow shortfalls, which proved critical during the rush for liquidity. Interestingly, some U.S. banks were shown to have particularly low measures of high-quality liquidity, as they were using assets such as tradable mortgage-backed securities as part of their liquidity buffers.
5. During a liquidity crisis, access to stable funding is key to survival; Canadian banks had a high ratio of retail to wholesale deposits. Retail deposits are insured and hence “sticky,” and provide a stable source of long-term funds for banks. In contrast, wholesale funds can withdraw rapidly upon minor negative news, and were a major source of
1 Prepared by L. Ratnovski and R. Huang.
3
vulnerability during the turmoil. The relative abundance of retail deposits seems to have been key for the resilience of Canadian banks.2 Easy access to retail deposits in Canada is explained by limited competition for household savings from non-banks, and by restrained bank asset growth that limited overall demand for funding.
6. Multivariate regression analysis confirms and extends these preliminary findings (Table 2). A high share of depository funding and a capital ratio above a critical minimum (although not the capital ratio per se) appear to be the most significant and robust determinants of bank resilience during the turmoil. Balance-sheet liquidity is less robust: it correlates with major failures, but not equity value declines (except very large ones). By considering interactions, the regressions also identify substitution between bank funding structure and capital, where, for a given risk profile, a bank with more depository funding can operate with lower capital, and a bank with higher capital can use less depository funding. In addition, larger banks have a higher probability of government intervention. However, rapid balance sheet expansion before the crisis appears irrelevant for performance during the crisis.
7. Regulatory and structural factors contributed to the resilience of Canadian banks by reducing their incentives to take risks. Canadian capital requirements are significantly more stringent than Basel minima (national targets of 7 percent for tier 1 capital and 10 percent for total capital, versus 4 and 8 percent prescribed by the Basel Accord). Banks are also subject to a maximum assets-to-total-capital multiple of 20 (corresponding to a leverage ratio of 5 percent). Besides providing an enhanced cushion, stringent capital requirements have beneficial incentive effects: they impede rapid balance sheet growth, restrict wholesale activities, and limit foreign expansion to niches where banks have clear competitive advantage not related to low cost of capital. Notable structural factors in Canada include high franchise values, a mortgage market characterized by prudent underwriting, and an overall prudent and conservative culture in the financial sector. Limited exposure to U.S. assets was a key additional factor behind the resilience of Canadian banks to the crisis.
Conclusions and policy implications
8. Canadian banks appear well positioned to weather the turmoil. A combination of strong capital and robust funding, in the context of sound regulation and supervision, has lent resilience to the banking system. With a severe recession underway, credit losses are likely to continue to climb, particularly on exposures to highly-leveraged households. But with banks stable and macroeconomic policies supportive, financial instability appears to be a tail risk.
2 During the crisis, Canadian banks have also had access to official funding programs such as the Insured Mortgage Purchase Program (IMPP), where the Department of Finance purchased already government-insured mortgages from banks to ease liquidity strains.
4
Pre-crisis (end-2006)
Bank Country Capital 1/ Equity decline Government intervention
1 Hypo Real Estate Holding AG GERMANY 2.1 97 Asset guarantees and public loans2 Deutsche Bank AG GERMANY 2.1 813 UBS AG SWITZERLAN 2.3 79 Asset guarantees4 Commerzbank AG GERMANY 2.5 89 Capital injection5 ABN Amro Holding NV NETHERLAND 2.6 ... Nationalized (carved out from Fortis)6 Barclays Plc UNITED 2.7 857 Fortis BELGIUM 2.8 94 Broken up, part nationalized8 Dresdner Bank AG GERMANY 3.0 ... Capital injection9 Northern Rock Plc UNITED 3.2 100 Nationalized
10 Dexia BELGIUM 3.3 89 Nationalized
20 Canadian Imperial Bank of Commerce CANADA 4.1 5421 Royal Bank of Canada RBC CANADA 4.3 4428 Banque de Montreal-Bank of Montreal CANADA 4.8 5329 Bank of Nova Scotia (The) - SCOTIABANK CANADA 4.9 4235 Royal Bank of Scotland Group Plc (The) UNITED 5.2 96 Capital injection, asset guarantees42 Toronto Dominion Bank CANADA 5.7 4350 Citigroup Inc USA 6.4 94 Recapitalized, asset guarantees63 Washington Mutual Inc. USA 8.5 100 Failed, taken over by FDIC64 JP Morgan Chase & Co. USA 8.6 5065 Bank of America Corporation USA 9.3 87 Capital injection, asset guarantees
Bank Country Liquidity 2/ Equity decline Government intervention
1 Capital One Financial Corporation USA 3.7 802 National City Corporation USA 4.0 100 Acquired by PNC Bank3 Citizens Financial Group Inc. USA 4.3 ... Not available (owned by RBS)4 SunTrust Banks, Inc. USA 4.3 855 US Bancorp USA 4.4 586 Washington Mutual Inc. USA 4.8 100 Failed, taken over by FDIC7 Regions Financial Corporation USA 5.0 908 Nomura Holdings Inc JAPAN 5.6 769 Wells Fargo & Company USA 6.0 47
10 Northern Rock Plc UNITED 6.7 100 Nationalized
41 Banque de Montreal-Bank of Montreal CANADA 23.99 5344 Toronto Dominion Bank CANADA 24.37 4345 Bank of Nova Scotia (The) - SCOTIABANK CANADA 24.43 4247 Royal Bank of Scotland Group Plc (The) UNITED 25.11 96 Capital injection, asset guarantees49 Bank of America Corporation USA 25.59 87 Capital injection, asset guarantees50 Canadian Imperial Bank of Commerce CANADA 26.00 5456 Royal Bank of Canada RBC CANADA 32.11 4463 Citigroup Inc USA 39.46 94 Capital injection, asset guarantees68 JP Morgan Chase & Co. USA 46.80 50
Bank Country Depository funding 3/ Equity decline Government intervention
1 Hypo Real Estate Holding AG GERMANY 24.0 97 Asset guarantees and public loans2 Northern Rock Plc UNITED 28.7 100 Nationalized3 Deutsche Bank AG GERMANY 34.1 814 BNP Paribas FRANCE 36.7 655 Citigroup Inc USA 37.8 94 Capital injection, asset guarantees6 HBOS Plc UNITED 41.0 100 Capital injection (part of Lloyds)7 Société Générale FRANCE 42.0 748 Banca Monte dei Paschi di Siena SpA ITALY 44.1 689 Dexia BELGIUM 44.9 89 Nationalized
10 DnB Nor ASA NORWAY 45.4 74
13 JP Morgan Chase & Co. USA 47.3 5015 Bank of America Corporation USA 47.9 87 Capital injection, asset guarantees33 Royal Bank of Scotland Group Plc (The) UNITED 59.3 96 Capital injection, asset guarantees51 Royal Bank of Canada RBC CANADA 65.1 4452 Banque de Montreal-Bank of Montreal CANADA 65.2 5357 Toronto Dominion Bank CANADA 67.9 4360 Canadian Imperial Bank of Commerce CANADA 68.2 5464 Bank of Nova Scotia (The) - SCOTIABANK CANADA 71.4 4269 Washington Mutual Inc. USA 74.6 100 Failed, taken over by FDIC
Sources: BankScope and staff calculations.1/ Equity over total assets >85% Due to an imminent failure2/ Liquid assets over total debt liabilities >70% Due to a severe deterioration3/ Depository funding over total assets
Table 1. Bank Fundamentals and Performance during Turmoil
Ten most vulnerable
Ten most vulnerable
(Jan 2007-Jan 2009)Bank performance during the turmoil
Selected banks
Selected banks
Ten most vulnerable
Selected banks
5
Cap
ital R
atio
-0.1
80.
26-3
.04
1.24
3.58
-31.
80-0
.91
3.68
-51.
0311
.88
190.
64-1
447.
58(-
0.42
)(0
.99)
(-1.
62)
(0.5
4)(1
.39)
(-2.
10)*
*(-
0.33
)(1
.31)
(-2.3
0)**
(0.1
3)(2
.31)
**(-
3.70
)***
Cap
ital R
atio
<4%
0.08
0.36
0.56
22.2
0(1
.58)
(1.7
1)*
(2.8
0)**
*(3
.85)
***
Bala
nce
Shee
t Liq
uidi
ty-0
.54
-0.4
3-0
.38
-1.5
3-1
.69
-1.1
9-0
.89
-1.1
5-0
.40
-33.
20-2
5.77
-17.
73(-
2.86
)***
(-3.
04)*
**(-
2.55
)**
(-2.4
0)**
(-2.
32)*
*(-
1.76
)*(-
1.35
)(-
1.31
)(-0
.55)
(-1.7
2)*
(-1.
32)
(-0.
99)
Dep
osito
ry fu
ndin
g-0
.33
-0.2
8-0
.48
-1.1
30.
54-3
.84
-1.1
61.
78-5
.54
-51.
6817
.66
-167
.95
(-2.
34)*
*(-
1.77
)*(-
2.80
)***
(-2.3
7)**
(0.7
6)(-
2.87
)***
(-2.
03)*
*(2
.03)
**(-2
.48)
**(-3
.16)
***
(0.7
1)(-
5.43
)***
Dep
osito
ry fu
ndin
g <5
0%-0
.01
0.47
0.58
13.0
7(-
0.54
)(2
.03)
**(2
.95)
***
(2.1
2)**
Cap
ital R
atio
* D
epos
itory
Fun
ding
5.21
58.1
587
.42
2545
.14
(1.7
2)*
(2.2
4)**
(2.2
5)**
(3.9
1)**
*
Log
(Ass
et)
0.06
0.05
0.05
0.14
0.14
0.17
0.12
0.09
0.13
4.02
2.35
4.99
(2.9
6)**
*(3
.09)
***
(2.7
2)**
*(1
.55)
(1.3
5)(1
.73)
*(1
.23)
(0.7
7)(1
.25)
(1.3
0)(0
.70)
(1.6
0)
N72
7272
6262
6262
6262
6262
62R
-Squ
ared
0.39
0.44
0.45
0.14
0.24
0.23
0.08
0.28
0.17
0.15
0.32
0.27
Sour
ce: S
taff
calc
ulat
ions
.
Tabl
e 2:
Ban
k Fu
ndam
enta
ls a
nd P
erfo
rman
ce: M
ultiv
aria
te R
egre
ssio
n R
esul
ts
Valu
e de
clin
e>70
%Va
lue
decl
ine
(%)
Not
es: A
sam
ple
of la
rge
OEC
D b
anks
(ass
ets
abov
e 10
0 bi
llion
euro
at e
nd-2
006)
. The
dep
ende
nt v
aria
bles
are
thre
e du
mm
y va
riabl
es fo
r im
min
ent b
ank
failu
re,
stoc
k m
arke
t val
ue d
eclin
e by
>85
%, a
nd s
tock
mar
ket d
ecin
e by
>70
%, a
s w
ell a
s th
e ab
solu
te v
alue
of s
tock
mar
ket d
eclin
e (s
tock
dec
line
mea
sure
d Ja
nuar
y 20
07 to
Jan
uary
200
9). C
apita
l rat
io is
the
equi
ty-to
-ass
ets
ratio
; bal
ance
she
et li
quid
ity is
the
liqui
d as
sets
-to-d
ebt-l
iabi
litie
s ra
tio; d
epos
itory
fund
ing
is th
e de
posi
ts-
to-a
sset
s ra
tio; a
ll ta
ken
at e
nd-2
006.
The
regr
essi
ons
invo
lvin
g du
mm
y va
riabl
es a
re e
stim
ated
bas
ed o
n Pr
obit,
with
coe
ffici
ents
tran
sfor
med
to b
e in
terp
rete
d as
pr
obab
ility
cha
nge
(0 to
1).
T-st
atis
tics
are
repo
rted
in p
aren
thes
is. *
**, *
*, * i
ndic
ate
stat
istic
al s
igni
fican
ce a
t the
1%
, 5%
, and
10%
leve
l.
Imm
inen
t fai
lure
Valu
e D
eclin
e>85
%
6
II. CANADIAN RESIDENTIAL MORTGAGE MARKETS: BORING BUT EFFECTIVE?1
1. Canada’s financial system has often been criticized for being “too conservative” or “not dynamic enough”. Indeed, when compared to the United States, Canadian banks seem to offer fewer loan options, in particular in the residential mortgage area. This could mean that households are underserved and that there is wide room for welfare improvements via increased financial innovation. However, Klyuev (2008) concluded that this is not the case and housing finance is highly advanced and sophisticated in Canada. Nevertheless, the same paper finds that financing options were somewhat limited, particularly at terms longer than five years.
2. This chapter concurs with previous research documenting the sophistication of Canada’s financial system, but suggests that regulations have limited the supply of some products. In particular, the paucity of longer-term loans is caused by a five-year maturity cap on government-guaranteed deposit insurance, and a prepayment penalty limit on residential mortgage loans in the Interest Act. The chapter also argues that for prime borrowers, the availability and cost of residential mortgages are comparable to those in the United States.
3. The Canadian predominance of shorter terms is driven by the more important role (versus in the United States) of retail deposits to fund mortgages, a feature driven by regulation. Deposits longer than five years are not popular because Canadian Deposit Insurance Corporation (CDIC) guarantees do not cover longer terms. Hence, Canadian banks have no natural funding for cost effective longer-term mortgages.
4. Also, regulations cause lenders to pass on the higher cost of hedging prepayment risk for longer mortgages in the form of higher interest rates (Figure 1). Section 10 of Canada’s Interest Act effectively fixes the prepayment penalty for most mortgages with a term to maturity greater than five years at three-months of interest, which is likely less than the penalty charged during the first five years of mortgage terms. Offsetting this to some degree is the portability of Canadian mortgages.2
5. Despite different regulations, mortgage costs are broadly similar in the United States and Canada. Even though at first sight mortgage rates are higher in Canada than in the United States, “posted” rates overstate actual transacted rates in Canada. Canadian five-year conventional rates have averaged about 100 basis points above the U.S. thirty-year conforming rate (Figure 2, in which both rates are normalized by their respective interest rate swap comparators).3 However, the Canadian rates are “posted” rates that overstate actual 1 Prepared by J. Kiff, based on a forthcoming IMF working paper. 2 U.S. homeowners that relocate must prepay their existing mortgages and take on a new one at prevailing rates. 3 Direct comparisons of fixed-rate mortgage costs are complicated by the fact that the term of “long-term” mortgage in Canada is five years, while it is thirty years or more in the United States. Comparing variable- or adjustable-rate mortgage (VRM or ARM) costs is complicated by the fact that U.S. ARMs embed numerous bells and whistles, such as “teaser rates” (see Kiff and Mills, 2007).
7
transacted rates, typically by more than 100 basis points. The Canadian Association of Accredited Mortgage Professionals (CAAMP) estimates that, on average, recent posted rates have exceeded transacted rates by 159 basis points (CAAMP, 2008).
6. In addition, the apparently lower U.S. thirty-year conforming rates reflect the payment of upfront points, which effectively prepay interest. For example, on February 19, 2009, the posted conforming rate was 5.04 percent with 0.7 points upfront, which is equivalent to 5.34 percent (plus 30 bps) with zero points. Moreover, there is an incremental term premium embedded in U.S. rates, which reflect the longer term of U.S. loans.
7. Anecdotal evidence suggests that the non-interest costs of originating and refinancing mortgage loans is significantly cheaper in Canada, which would also contribute to closing any apparent gap between the costs in both countries. Canadian borrowers pay about C$2,000 in upfront fees and taxes for a new loan, and on a refinancing about C$1,000 plus a prepayment penalty of about C$3,000 on the old mortgage.4 On the same loans (new loans and refinancings), U.S. borrowers pay origination fees of $1,000 to $3,000, plus about $1,000 of costs and fees, and local government taxes of about $1,000.
8. Payment affordability criteria for prime borrowers are broadly similar in both countries. For example, in order to qualify for mortgage insurance in Canada, gross debt service should usually not exceed 32 percent of gross household income, and total debt service cost should usually not exceed 40 percent (versus 28 and 36 percent to qualify for Fannie Mae and Freddie Mac insurance).
9. However, the approval criteria for adjustable-rate loans in Canada are usually based on the three-year fixed-term rate, which is usually the highest fixed rate inside of the five-year term, whereas U.S. practice is to use the current floating rate.5 Canada also has a small “Alt-A” market aimed mainly at self-employed people who have difficulty documenting their stated income.
10. Down payment requirements are roughly in line with those in the United States. Canadian federally-regulated deposit-taking institutions have been able to underwrite insured mortgages with loan-to-value ratios as high as 95 percent since 1992, and occasionally before then.
11. There are no limits to the size of individual loans that the Canada Mortgage and Housing Corporation (CMHC) and other mortgage insurers will insure, which minimize risks to banks’ balance sheets. In the United States, Fannie Mae and Freddie Mac 4 Cost calculations are based on a $240,000, five-percent loan and based on transactions in Ottawa, Ontario (provided by Steven Sheppard of BrazeauSeller LLP) and McLean, Virginia (AimLoan.com). 5 In fact, until recently, it was U.S. practice to use a fixed “teaser rate” that applied to the first two or three years of many adjustable-rate mortgages (ARMs), for affordability calculations (Kiff and Mills, 2007). However, some Canadian lenders have started to qualify adjustable-rate loans on the basis of current floating-rate loan rates.
8
insurance is only available on loans up to the “conforming limit”, which vary by geographic areas, but in 2009 is $625,500 for loans on single-family homes in “high-cost” areas.
Conclusions and Policy Implications 12. The availability and costs of Canadian residential mortgage loans to prime borrowers are comparable to those in the United States. Moreover, even though there are clear institutional differences, homeownership in both countries is virtually identical at about 68 percent of all households. This said, some aspects of Canada’s mortgage market can be improved. To encourage the development of longer-term mortgage markets, the government might consider dropping the five-year cap on CDIC deposit insurance and the five-year prepayment penalty fixed in the Interest Act. Until that happens, rates on fixed-term residential mortgages beyond the five year term will remain uneconomical for most borrowers. The opening up of longer fixed-rate terms would help households to better manage financial risks.
Figure 1: Canadian Bank Mortgage Rates (February 20, 2009)(percent)
0
1
2
3
4
5
6
7
8
1 2 3 4 5 6 7 8 9 10
BMO TDRBC CIBCBNS Swap
(years of maturity) Figure 2: Residential Mortgage versus Interest Rate Swap Rates
(basis points)
0
50
100
150
200
250
300
350
400
450
500
Sep-99 Aug-00 Jul-01 Jun-02 May-03 Apr-04 Mar-05 Feb-06 Jan-07 Dec-07 Nov-08
Canadian 5Y Conventional v 5Y Swaps
U.S. 30Y Conforming v 10Y Swaps
9
III. THE IMPACT OF GLOBAL SHOCKS ON CANADA—WHAT DO MACRO-FINANCIAL LINKAGES TELL US?1
1. The paper analyzes how Canada’s strong macro-financial ties with the United States affect its real activity. It builds on related literature,2 by focusing on the role of financial conditions in driving Canadian growth. For instance, in the current global crisis, Canada has been affected not only by declining external demand but also by global financial shocks, the latter inducing domestic financial strains. Notably, credit conditions have remained tight notwithstanding aggressive monetary policy rate cuts. This paper sheds light on how these external and domestic shocks bear on real GDP growth.
2. These issues are assessed by building a macro-financial conditions index for Canada using a Bayesian VAR (BVAR) estimation model, which uses “informative priors” about the steady-state values of the variables to limit loss in estimation precision caused by the heavy parameterization of VARs.3 This is especially important given the short sample size, as data on Canadian banks’ lending standards begin in 1999. The model is then used for scenario analyses—i.e., projecting Canada’s growth outlook for alternative external and domestic conditions—and building a macro-financial index for growth. The “macro” part of the index accounts for the impact of external demand conditions, while the “financial” part accounts for external and domestic financial conditions.
3. Banks’ non-price lending standards in the United States and Canada are used to measure credit availability for Canadian businesses. Canadian lending standards (from the Bank of Canada’s Senior Loan Officer’s Survey (SLOS)) are strongly correlated with U.S. SLOS for large corporations (Figure 1). Both measures are negatively correlated with Canada’s contemporaneous and future real GDP growth, which is not surprising given that Canadian businesses raise one-quarter of their funding in the United States.4 For instance, the
1 Prepared by R. Duttagupta and N. Barrera.
2 See Swiston and Bayoumi (2008), IMF WP/08/03, and Klyuev (2008), IMF WP/08/23.
3 The BVAR model assumes that the forecaster has potentially useful information on the steady state values of the variables used in the model, i.e., an informative prior would make forecasts converge to a level that the forecaster judges reasonable. If the forecaster is correct, this leads to a substantial improvement in the forecasting performance of BVAR compared with a standard VAR.
4 See Klyuev (2008), IMF WP/08/22.
-60
-40
-20
0
20
40
60
80
100
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008-6
-4
-2
0
2
4
6
8
10
Source: Haver Analytics.
Figure 1. Banks' Lending Standards for Corporations and Real GDP Growth
Non-price lending standards for large businesses (Percentage of respondents that tightened lending standards minus the percentage that eased lending standards, left axis)
United States Canada
Real GDP growth(year-on-year)
(Percentage of respondents that tightened lending standards minus the percentage that eased lending standards)
10
-1.6
-2.6
-1.9
-2.9
-4
-3
-2
-1
0
1
2
Unconditional Forecast Forecast conditional on US growth 1/
US real GDP growth, 2009
Canada real GDP growth, 2009
1/ The path for US real GDP is assumed to be 1 percentage point lower than the unconditional forecast.
Figure 3. Canadian Real GDP Forecast for 2009 Under Alternative paths for US Real GDP Growth(percent)
Figure 4. Canadian Real GDP Forecast Under Alternative Conditions for Canadian SLOS
-1.9 -2.3
35
45
-10
0
10
20
30
40
50
60
Unconditional Forecast Forecast conditional on Canadian SLOS 1/
Canada real GDP growth, 2009 (percent)
Canada average SLOS, 2009 (percentage of respondents that tightenedlending standards minus the percentage that eased lending standards)
1/ The path for Canadian SLOS is assumed to be 10 percentage point tighter than the unconditional forecast.
-0.40
-0.35
-0.30
-0.25
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0 1 2 3 4 5 6 7 8 9 10 11 12
US SLOSCanadian SLOS
US high yield spread
(quarters)
(percentage points)
Figure 2. Impulse Response Functions(Impact on Canada's real GDP growth of one standard deviation shocks to financial indicators)
deceleration in Canadian growth in the early 2000s and more recently coincided with a sharp tightening of one or both measures of lending standards.
4. The baseline model comprises nine macro-financial variables: (i) U.S. real GDP growth (expressed as year-over-year percent change), (ii) oil price growth, (iii) U.S. three-month t-bill rate, as a proxy for U.S. monetary conditions, (iv) U.S. SLOS, (v) U.S. high-yield spread over the 10-year treasury yield, (vi) Canada’s real GDP growth, (vii) Canadian SLOS, (viii) Canada’s three-month t-bill rate, and (ix) growth in Canada’s real effective exchange rate. The U.S. variables are allowed to be exogenous with respect to the Canadian variables. The data cover the period 1999Q2−2008Q4 at a quarterly frequency. While the steady-state priors were drawn from existing literature or historical averages, the posterior values from the model are close to the priors, implying that the chosen priors are reasonable.
5. Impulse-response functions confirm that financial shocks have a significant effect on growth (Figure 2). A one-percent standard deviation (s.d) shock to the U.S. SLOS—a net tightening of 9.5 percentage points (pp)—reduces Canadian growth by about 0.4 pp in six quarters, while the impact of a similar tightening in the Canadian SLOS is weaker but works with a shorter lag—reducing growth by ¼ pp in three quarters. The larger effect on growth of a tightening in the US SLOS than the Canada SLOS could reflect indirect effects of a higher U.S. SLOS on Canada—tighter U.S. lending standards slow U.S. growth (and hence external demand) while also spilling over to Canadian credit conditions. Also, a one s.d. shock to the U.S. high-yield spread (129 basis points) shaves off 0.15 pp from growth over three quarters.
6. Looking ahead, Canada’s economic outlook would be firmly tied to prospects for external activity as well as also domestic financial conditions. Conditional forecasts show that if U.S. growth in 2009 were 1 pp lower than that projected by the unconditional model, then Canada’s growth would also be 1 pp lower (Figure 3), while if Canadian SLOS were 10 pp tighter in 2009, then Canada’s growth would be 0.4 pp lower (Figure 4).
11
7. The variance decompositions for Canada’s real GDP growth confirm that foreign shocks are the most important source of variation in Canada’s growth over the long run, with U.S. growth accounting for close to 42 percent, U.S. financial shocks another 28 percent, and oil prices 14 percent (Figure 5). The contribution of domestic financial conditions to Canadian growth increases from 0 percent in the short-term to over 8 percent in 12 quarters. Oddly enough, the contribution of Canadian growth to its own variance declines from 43 percent to a little over 7 percent in the long run—this could reflect the fact that the model is based on a recent sample period (since 1999), when the openness of the Canadian economy to external volatilities, especially vis-à-vis the United States has increased markedly. Indeed, a simple monetary BVAR model estimated starting in the early 1990s (which excludes data on financial market indicators, i.e., lending standards and high yield spreads) attributes a larger role to domestic growth shocks.
8. A macro-financial condition index built from the coefficients of the baseline model tracks real GDP growth well, and shows that tightening in the Canadian SLOS and effects of past real appreciation have played a key role in the deceleration of the Canadian growth rate in the run up to the recent crisis. However, U.S. economic and financial conditions will increasingly bear on growth in the near term.
Conclusions and policy implications
9. The paper underscores the need for a normalization of U.S. and domestic financial strains for Canada’s recovery. Tight U.S. and domestic financial conditions depress Canada’s growth by intensifying credit strains for businesses, although monetary easing helps ameliorate these strains somewhat. Thus, stability in U.S. financial conditions would be critical for a sustained pick up in Canadian economic activity.
Figure 6. Contribution to the Macro-Financial Conditions Index
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2001 2002 2003 2004 2005 2006 2007 2008 2009
US GDP Oil US SLOS
US tbill US HYS CAN SLOS
Can tbill CAN REER FCI
(Percentage points of real GDP growth, 4Q moving average)
Figure 5. Variance Decomposition of Canadian Real GDP Growth(percent)
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8 9 10 11 12
Domestic financial factors
US real GDP
Oil
Canadian real GDP US financial factors
(quarters)
12
IV. IS THE CANADIAN HOUSING MARKET OVERVALUED? A TALE OF TWO REGIONS1
1. Canadian house prices have increased significantly between 2003 and early 2008, with a marked downward trend since mid-2008 (Figure 1). House prices rose by around 60 percent in nominal terms (45 percent in real terms) from 2003 to the peak, before falling by around 10 percent (both in real and nominal terms) in the latter part of 2008. The decline is particularly acute in the west given the collapse in commodity prices, although modest declines are occuring elsewhere.
2. This chapter summarizes estimates of the gap between actual house prices and their equilibrium levels for five large Canadian provinces (Alberta, British Columbia, Ontario, Quebec, and Saskatchewan). The price measure used is the existing home price from the Canadian Real Estate Association's Multiple Listing Service (MLS) database, and is deflated by each province’s CPI.2 We examine current valuations against economic fundamentals using quarterly regional data—such as disposable income, demographic developments, and mortgage credit. The analysis is based on an error correction model, which combines the long-run, cointegrating relationship among the levels variables and the short-run relationships among the first differences of the variables.
3. The error correction model postulates that the growth rate of real houses is explained by a combination of the following factors (depending on the province considered) (Table 1):3
o Past growth rates of real house prices. For most provinces, we find that the current growth rate is positively correlated with the past growth rate.
1 Prepared by E. Tsounta.
2 While private banks and other forecasters have recently developed new indices on house prices (e.g., adjusting for quality), CREA’s sales weighted index remains the most widely used, including by Canada’s Mortgage and Housing Corporation and Finance Canada, as a major economic indicator given its larger sample size (all provinces, more years, all sales by realtors). This measure exhibits the largest volatility, including large upswings, and in that respect it should represent an upper limit in terms of any overvaluation. For example, the quality-adjusting Teranet-National Bank house price index (data starting in 1999) records that house prices in Canada’s six metropolitan areas of Ottawa, Toronto, Calgary, Vancouver, Montreal, and Halifa have risen on average by 48 percent from 2003 to their peak, versus around 60 percent using CREA’s estimates.
3 There is considerable uncertainty about the right technique to model equilibrium house prices. Papers that cite limitations in identifying the determinants of home prices include Allen, J. et al. (2006), Klyuev (2008), and IMF (2004).
60
80
100
120
140
160
180
200
220
Mar-90 Mar-92 Mar-94 Mar-96 Mar-98 Mar-00 Mar-02 Mar-04 Mar-06 Mar-08
CanadaQuebecOntarioSaskatchewanAlbertaBritish ColumbiaOther
Figure 1. Real House Prices (1990=100, NSA)
Source: CREA, Finance Canada, and Haver Analytics.
13
o Reversion to fundamentals implied by the long-run equation. We find that only for Ontario, the growth rate of house prices shows long-run reversion to the equilibrium prices derived from the model, implying that prices would tend to fall when they are out of line with fundamentals.
o Economic fundamentals. For most provinces, we find that the growth rate of house prices is positively affected by (per capita) real income growth—as this increases households’ purchasing power and borrowing capacity—and positively affected by mortgage credit growth (higher rates indicate that households are less credit rationed), and population growth (as a proxy for the growth rate of households).
A lbe rta B rit ish Co lum b ia O nt ario Q uebec S as ka t chew an
E xp lan ator y V ari abl es
Lagged dependen t v ariab le 0 .43 -0 .08 -0. 33 -0 .64 0.03 Lagged rea l house price (growth ) [3 .32 ] [-0 .6 ] [-3 . 51] [-4 . 1 ] [ 0 .22 ]
R evers ion -0 .03 0. 07 -0. 64 0 .02 -0. 01 E rro r co rrec ti on c oe ff icien t [ -0 .87 ] [1 .8 ] [-6 . 81] [1 .1 ] [-0 .07 ]
F undamen ta ls
Real ea rn ings (pe r capita , g rowth ) 0 .89 1 .6 0 .56 -0 .07 0.96[2 .96 ] [3 .4 ] [1 .38 ] [ -0 .13 ] [ 2 .16 ]
R ea l c red it (g rowth ) 0 .09 -0. 05[1 .89 ] [-1 . 16]
P opula t ion (g rowth ) -5 .9 -18 .14 9.85[-1 .7 ] [-3 . 77] [ 2 .88 ]
S ourc e : S ta f f es tim a tes . N o te: T -st a tis tic s are list ed w ith in b rack e t s.
D ependen t V a riab le : R ea l H ous e P ric e (g row th )
Tab le 1 . De t ermi nan ts o f H ouse P rices in S e lec ted Canad ian P rovi nces(S um mary of E m p irica l Re su lt s, 1992-2008)
Conclusions and policy implications
4. Results reveal that prices in the west are above the levels implied by the model, although prices remain close to or slightly below equilibrium in the east (Figure 2). Indeed, the econometric model indicates that most of the recent surge in Canadian home prices, even in the west, reflects a catch-up from stubbornly undervalued levels following the housing collapse in the early 1990s, rather than a housing bubble per se.4 While resource-rich western provinces continue to have house prices above the model prediction, their prices have diminished significantly in the last year. In contrast, Quebec and Ontario appear to be close to equilibrium, or slightly below the prices implied by the model, indicating a divergence between western and eastern provinces in house price dynamics.
4 IMF (2004, 2009) reaches similar conclusions in a cross-country analysis.
14
5. While a crash in the national housing market appears unlikely, a correction in western housing markets could have national implications. With the west accounting for 35 percent of Canada’s GDP and around 30 percent of Canada’s labor force, an abrupt correction in its housing market, could cause adverse spillovers to the rest of Canada; with housing assets and mortgage debt at record ratios of disposable income, household balance sheets are particularly exposed to house price dynamics. Similarly, given that most mortgages are originated by banks (55 percent of which do not carry mortgage insurance but have a loan-to-value ratio below 80 percent),5 this could also somewhat impact the banking sector, affecting the future provision of mortgages and credit in general, imposing additional downward pressure on spending, incomes, wealth and thus house prices. Last but not least, house prices directly impact headline and core inflation, thus affecting inflationary expectations as well.
5 Chapter II discusses Canadian mortgage markets.
15
Figure 2. Canada: House Price Over/Undervaluation(In percent)
Sources: Finance Canada, CREA, Fund staff calculations.
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
3
1993 1994 1996 1997 1999 2000 2002 2003 2005 2006 2008
Quebec
-20
-15
-10
-5
0
5
10
15
20
25
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Saskatchewan
-20
-15
-10
-5
0
5
10
15
20
1991 1993 1994 1996 1997 1999 2000 2002 2003 2005 2006 2008
Alberta
-6
-4
-2
0
2
4
6
8
10
12
1995 1996 1998 1999 2001 2002 2004 2005 2007
Ontario
-20
-15
-10
-5
0
5
10
15
20
1992 1994 1995 1997 1998 2000 2001 2003 2004 2006 2007
British Columbia
16
V. HOW DO COMMODITY PRICES AFFECT ECONOMIC SLUMPS IN COMMODITY EXPORTERS? 1
1. The financial crisis has been accompanied by a sharp fall in commodity prices, with deleterious effects on commodity exporters. Commodity exporters, including Canada, have been subject to twin shocks during the ongoing crisis—on external demand and on commodity prices (Figure 1). The latter could affect the duration and severity of Canada’s ongoing recession, but current research sheds scant light on the relationship between commodity prices and business cycle characteristics in commodity exporting countries.2
2. This chapter analyzes whether the length and depth of recessions in commodity exporters is affected by the commodity-price cycle. Commodity prices can have offsetting effects on business cycles of commodity exporters. On the one hand, commodity booms raise real incomes and support domestic demand. On the other hand, an appreciation of the real exchange rate in response to the upturn in commodity prices can affect external competitiveness and stifle real exports, depressing external demand. Both of these effects were experienced by Canada through the first half of 2008. This analysis takes into account these offsetting effects.
3. The sample comprises 15 advanced and emerging market countries with varying degrees of dependence on commodity exports (Table 1). While commodities comprise close to 10 percent of total exports in most of these countries, ten are net commodity exporters—Australia, Brazil, Canada, Chile, Denmark, Mexico, New Zealand, Norway, Russia, and South Africa. The other five—Netherlands, Sweden, Switzerland, the United Kingdom, and the United States—form a control group of net commodity importers. In the 1 Prepared by R. Duttagupta.
2 Claessens, Kose and Terrones (WP/08/274) analyze linkages between key macroeconomic and financial variables around business cycles for a large number of OECD countries including commodity exporters, focusing on the roles of credit, house and equity prices.
Commodity Exporters
Net Commodity Exports
(Percent of Total Exports)
Australia 41.8Brazil 25.5Canada 15.7Chile 39.5Denmark 5.5Mexico 3.1New Zealand 17.3Norway 64.0Russia 54.7South Africa 0.5
Commodity ImportersNetherlands -1.0Sweden -2.9Switzerland -5.3The United Kingdom -2.1The United States -22.8
Source: International Monetary Fund (WEO)
Table 1. Sample of Countries
Figure 1. Commodity Price Index (2005 = 100) 1/
20
40
60
80
100
120
140
160
180
200
Mar-81 Mar-84 Mar-87 Mar-90 Mar-93 Mar-96 Mar-99 Mar-02 Mar-05 Mar-08
Source, World Economic Outlook, Statistics Canada1/ World price expressed as simple average of the following five categories: food, agriculture, metals, gold, and crude oil.
Canada
World
17
absence of data for country-specific commodity price indices for most of the sample countries, commodity export prices are constructed as weighted averages of the global commodity price indices (food, agriculture, crude oil, gold, and metals), where the weights are given by the export share of each commodity in the total commodity exports of each country (2002−04 averages), provided by the IMF’s World Economic Outlook.
4. The sample includes 120 spells of economic downturns during 1981−2008 with considerable differences in experiences across the countries. An economic downturn is defined as the period from peak to trough in (year-on-year) real GDP growth. The duration of the downturn is measured in quarters, and a spell is considered only if it lasted at least three quarters. The data do not reveal any particular trend in the nature of downturns (Figure 2)—while the average length of a spell was about 5½ quarters, some commodity exporters experienced longer recessions averaging 6½ quarters (New Zealand, Norway, South Africa). At the same time, Sweden, a net commodity importer, experienced the longest downturns on average (7 quarters), while Brazil and the United States experienced the shortest (4.2 quarters). For commodity exporters, almost all spells were associated with a fall in commodity prices—either in the run up or during the downturn—signaling their importance in explaining business-cycle fluctuations in these economies.
5. Estimation results confirm that large declines in commodity prices significantly exacerbate ensuing economic downturns in commodity exporters. A fall in pre-crisis commodity prices induces longer and sharper recessions; the latter effect measured by the difference between peak and trough real GDP growth rates.
• Recessions are relatively longer for commodity exporters that experience a pre-crisis fall in commodity prices (Table 2, Column A). In particular, a 10 percent decline in commodity prices prior to the downturn increases the length of the recession spell by 3/5th of a quarter, on average. However, this result is statistically insignificant for the overall sample including commodity importers.
Figure 2. Average length of Economic Downturns(in quarters)
3.0
4.0
5.0
6.0
7.0
8.0
Aust
ralia
Bra
zil
Can
ada
Chi
le
Den
mar
k
Mex
ico
New
Zea
land
Nor
way
Rus
sia
Sou
th A
frica
Net
herla
nds
Sw
eden
Sw
itzer
land
The
Uni
ted
Kin
gdom
The
Uni
ted
Sta
tes
18
• Commodity prices also affect the severity of the ensuing recession, whereby a 10 percent decline in pre-crisis commodity prices reduces real GDP growth at the trough by 0.4 percentage point, a result that is significant for both the overall and restricted samples (Table 2, Column B).
6. The results also show that longer and deeper downturns are positively associated with the pre-crisis real GDP growth. The higher the real GDP growth before the crisis, the longer it takes for the ensuing recession to “bottom out” and the lower is real growth at the bottom. The other results are not significant—while the regression also controls for the impact of the real exchange rate, the latter does not have a significant effect on influencing duration or depth of the economic downturn.
Overall sample
Restrictedsample 3/
Overall sample
Restrictedsample 3/
Explanatory Variables
Pre-slump real GDP growth 0.60(0.01)**
0.60(0.02)**
0.88(0.00)**
1.03(0.00)**
Commodity price growth -0.02(0.25)
-0.02(0.09)*
-0.03(0.02)**
-0.04(0.02)**
Merchandise export growth 0.01(0.75)
0.00(0.90)
0.02(0.16)
0.01(0.71)
REER appreciation -0.05(0.30)
-0.06(0.30)
-0.01(0.86)
-0.02(0.60)
1/ Economic Slump is length of spell from peak to trough. Explanatory variables represent levels in the quarter before the start of the slump. Growth rates are expressed in year-on-year percentage changes.2/ * and ** refer to significance at 5 percent and 10 percent levels (p-value in parentheses).3/ Includes net commodity exporters only.
Table 2. Fixed Effect Regression of Length and Depth of Economic Slump 1/
Coefficient
(A)Dependent variable: length of recession
(B)Dependent variable: depth of recession
Coefficient
Conclusions and policy implications
7. Commodities play a key role in affecting the severity and length of economic downturns in commodity exporters. The possible ameliorating effect of falling commodity prices on the real exchange rate and exports is likely more than offset by the negative effect on domestic incomes and activity. On average, a 10 percent decline in commodity prices (y-o-y) prolongs the ensuing downturn by 0.6 quarters and deepens it by 0.4 percentage point in real GDP growth. While these results do not indicate how much worse it could be for countries that continue to experience falling commodity prices after the downturn is underway, 3 they stress that a sustained rebound in commodity prices is crucial for Canada’s economic prospects.
3 Since the analysis eliminates reverse causality from the downturn to commodity prices by focusing on pre-crisis commodity price growth, the above results do not provide an answer for how falling commodity prices after the start of a downturn would affect the length and depth of the downturn.
19
References
Allen, Jason, Robert Amano, David. P. Byrne, and Allan W. Gregory, 2006, “Canadian City
Housing Prices and Urban Market Segmentation,” Working Paper 2006-49, Bank of Canada (Ottawa).
Canadian Association of Accredited Mortgage Professionals (CAAMP), 2008, Annual State
of the Residential Mortgage Market in Canada, November. Claessens, Stijn, M. Ayhan Kose, and Marco Terrones, 2008, “What Happens During
Recessions, Crunches, and Busts?,” IMF Working Paper 08/274 (Washington: International Monetary Fund).
International Monetary Fund, 2004, World Economic Outlook, September 2004: A Survey by
the Staff of the International Monetary Fund, World Economic and Financial Surveys (Washington).
_____, 2009, World Economic Outlook, April 2009: A Survey by the Staff of the
International Monetary Fund, World Economic and Financial Surveys (Washington). Kiff, John, and Paul Mills, 2007, “Money for Nothing and Checks for Free: Recent
Developments in U.S. Subprime Mortgage Markets,” IMF Working Paper 07/188 (Washington: International Monetary Fund).
Klyuev, Vladimir, 2008, “Show Me the Money: Access to Finance for Small Borrowers in Canada,” IMF Working Paper 08/22 (Washington: International Monetary Fund). _____, 2008, “Real Implications of Financial Linkages Between Canada and the United States,” IMF Working Paper 08/23 (Washington: International Monetary Fund). _____, 2008, “What Goes Up Must Come Down? House Price Dynamics in the United States,” IMF Working Paper 08/187 (Washington: International Monetary Fund). Swiston, Andrew, and Tamim Bayoumi, 2008, “Spillovers Across NAFTA,” IMF Working Paper 08/03 (Washington: International Monetary Fund).